Abstract: A method for rescue support involves distributing transponder devices to individuals in a geographic region and causing the transponder devices to be activated. If a large-scale emergency situation occurs, a fleet of unmanned aerial vehicles are operated within the geographic region to receive distress signals transmitted by the activated transponder devices. A computer system processes the distress signals to obtain detection data for the individuals, the detection data comprising a location of the respective transponder device and a health status of the respective individual. The computer system generates, based on the location and health status of the respective individual, a prioritization chart for the geographic region, the prioritization chart being indicative of one or more sub-regions to be prioritized for a rescue operation.

Abstract: Triangulation is applied, by a method and a device, to determine a scene point location in a global coordinate system, GCS, based on location coordinates for observed image points in plural views of a scene, and global-local transformation matrices for the views. The local coordinates are given in local coordinate systems, LCSs, arranged to define distance coordinate axes that are perpendicular to image planes of the views. The scene point location is determined by minimizing a plurality of differences between first and second estimates of the scene point location in the LCSs, the first estimates being given by linear scaling of the location coordinates, and the second estimate being given by operating the transformation matrices on the scene point location. An improved robustness to observation errors is achieved by defining the plurality of differences to penalize differences that includes the linear scaling as applied along the distance coordinate axes.

Abstract: A device determines the positioning of objects in a scene by implementing a robust and deterministic method. The device obtains object detection data (ODD) which identifies the objects and locations of reference points of the objects in views of the scene. The obtained ODD is processed to identify a first image object of a first view as a mirror reflection of a real object. A virtual view associated with a virtual camera position is created, including the ODD associated with the first image object of the first view. The ODD associated with the first image object is removed from the first view. Based on the ODD associated with at least said virtual view and a further view of the one or more views, a position of said first image object is computed.

Abstract: An electronic device includes a memory circuitry, an interface circuitry, a sensor circuitry, and a processor circuitry having a predictor circuitry configured to operate according to a prediction model. The processor circuitry is configured to obtain sensor data. The processor circuitry is configured to determine, based on the sensor data, using the predictor circuitry, a predicted time parameter indicative of a prediction of a time slot to initiate an establishment of a network with one or more other electronic devices. The processor circuitry is configured to initiate the establishment of the network with the one or more other electronic devices according to the predicted time parameter.

Abstract: A processing device is configured to obtain a sequence of images of a scene captured by an image sensor, determine an analysis area for an object in a respective image in the sequence of images, and process the respective image within the analysis area for detection of predefined features of the object. The processing device is further configured to receive pose prediction data, PPD, which represents predicted poses of the object as a function of time, and to determine the analysis area based on the PPD. The PPD may be given by three-dimensional poses of the object that have been determined in the system based on images from a plurality of image sensors in the system. The PPD facilitates detection of features of individual objects in the images even if the objects are occluded and/or crowded.

Abstract: A system for training a neural network for action recognition based on unlabeled action sequences includes a first neural network (NN1) and a second neural network (NN2). A first updating module is arranged to update parameters of NN1 to minimize a difference between representation data generated by NN1 and representation data generated by NN2. A second updating module is arranged to update parameters of NN2 as a function of the parameters of NN1. An augmentation module includes first and second sub-modules and is configured to include augmented versions of incoming action sequences in first and second input data. The first and second sub-modules are configured to apply at least partly different augmentation to the incoming action sequences. After NN1 and NN2 have been operated on one or more instances of the first and second input data, NN1 comprises a parameter definition of a pre-trained neural network.

Abstract: A device determines positions of objects in a scene. The device obtains object detection data (ODD) which identifies the objects and locations of reference points of the objects in 2D images of the scene. The device processes the ODD to generate candidate association data (CAD) which associates pairs of objects between the images, computes estimated 3D positions in the scene for associated pairs of objects in the CAD, and performs clustering of the estimated positions. The device further generates, based on estimated 3D positions in one or more clusters, final association data (FAD) which associates one or more objects between the images, and computes one or more final 3D positions in the scene for one or more reference points of the one or more objects in the FAD. The final 3D position(s) represent the 3D position or the 3D pose of the respective object in the scene.

Abstract: A method for extrinsic calibration of cameras in a monitoring system, performed by a calibration device, includes obtaining preliminary calibration data of the cameras, obtaining a preliminary pose for an object observed by the cameras, and determining calibration data that results in an optimization of an objective function using the preliminary calibration data and the preliminary pose as starting values for the optimization. The method further includes defining the objective function with calibration data and pose as unknown variables and including a first function yielding aggregated reprojection error, and a second function yielding pose quality based on relative orientation and relative distance between 3D points of the pose. The second function enables improved accuracy of the calibration data and may be configured to yield pose quality for individual poses or motion quality for a time sequence of poses.

Abstract: A device for training a first machine learning-based model (MLM) for action recognition implements a training method. According to the training method, the training device obtains training data that comprises time sequences of data samples, which represent predefined subjects that are performing predefined actions. The training device trains the first MLM based on the training data, to discriminate between the predefined actions and to be adversarial to discrimination between the predefined subjects by a second MLM, and trains the second MLM based on feature data that is extracted by the first MLM for the training data, to discriminate between the predefined subjects. Thereby, the first MLM is encouraged to extract feature data that is unrelated to individual subjects, which improves action recognition performance of the trained first MLM when encountering new subjects.

Abstract: A method for processing images of animals obtains first training data, which includes first images of animals and first annotation data including first location data for predefined feature points on at least one animal in each image in the first images, and trains a first machine learning-based model (MLM) by use of the first training data to generate a first trained MLM. The method obtains second images of animals, operates the first trained MLM on the second images to generate second annotation data comprising second location data for the predefined feature points on at least one animal in each image in the second images, and aggregates at least a subset of the first and second images, and a corresponding subset of the first and second annotation data, into second training data, which thereby is expanded in relation to the first training data.

Abstract: An electronic device includes memory circuitry, and processor circuitry having an action detection circuitry configured to operate according to an action detection model for detecting an action based on a machine-learning scheme. The processor circuitry being configured to obtain sensor data; generate, based on the sensor data, a set of features associated with a frame; determine, based on the set, using the action detection model, whether the frame corresponds to a sub-action; apply a nondeterministic finite automaton, NFA, scheme, to the determined sub-action for the frame, wherein the NFA scheme has a set of states associated with corresponding sub-actions and is configured to output one or more action classes; determine, using the NFA scheme, an action class; detect the action based on the action class; and increment an action counter based on the detected action.

Abstract: A device determines the positioning of objects in a scene by implementing a robust and deterministic method. The device obtains object detection data (ODD) which identifies the objects and locations of reference points of the objects in views of the scene. The obtained ODD is processed to identify a first image object of a first view as a mirror reflection of a real object. A virtual view associated with a virtual camera position is created, including the ODD associated with the first image object of the first view. The ODD associated with the first image object is removed from the first view. Based on the ODD associated with at least said virtual view and a further view of the one or more views, a position of said first image object is computed.

Abstract: An electronic device includes a memory circuitry, an interface circuitry, a sensor circuitry, and a processor circuitry having a predictor circuitry configured to operate according to a prediction model. The processor circuitry is configured to obtain sensor data. The processor circuitry is configured to determine, based on the sensor data, using the predictor circuitry, a predicted time parameter indicative of a prediction of a time slot to initiate an establishment of a network with one or more other electronic devices. The processor circuitry is configured to initiate the establishment of the network with the one or more other electronic devices according to the predicted time parameter.

Abstract: A method for extrinsic calibration of cameras in a monitoring system, performed by a calibration device, includes obtaining preliminary calibration data of the cameras, obtaining a preliminary pose for an object observed by the cameras, and determining calibration data that results in an optimization of an objective function using the preliminary calibration data and the preliminary pose as starting values for the optimization. The method further includes defining the objective function with calibration data and pose as unknown variables and including a first function yielding aggregated reprojection error, and a second function yielding pose quality based on relative orientation and relative distance between 3D points of the pose. The second function enables improved accuracy of the calibration data and may be configured to yield pose quality for individual poses or motion quality for a time sequence of poses.

Abstract: An electronic device is configured to analyze movement data generated by a sensor (e.g., an accelerometer assembly) in the electronic device to determine if the electronic device should be activated from a reduced power state. A control circuit in the electronic device can compare the movement data to a predetermined movement pattern to determine whether the movement should be ignored, or if the movement is consistent with a user shaking the electronic device to activate the electronic device. As a result of this approach, erroneous activation of the electronic device is avoided, thereby preventing the electronic device from needlessly depleting its battery before deployment by the end user.

Abstract: An object monitoring system includes a plurality of wireless sensor devices and at least one user device A method for controlling a response time of one of the wireless sensor devices in such a system includes determining one or more metrics representing a present state of an object; and calculating a probability that one of at least one user devices will interact with the wireless sensor device within a pre-determined time period by executing a trained model with the one or more determined metrics as input and controlling the response time of the wireless sensor device based on the calculated probability.

Abstract: Triangulation is applied, by a method and a device, to determine a scene point location in a global coordinate system, GCS, based on location coordinates for observed image points in plural views of a scene, and global-local transformation matrices for the views. The local coordinates are given in local coordinate systems, LCSs, arranged to define distance coordinate axes that are perpendicular to image planes of the views. The scene point location is determined by minimizing a plurality of differences between first and second estimates of the scene point location in the LCSs, the first estimates being given by linear scaling of the location coordinates, and the second estimate being given by operating the transformation matrices on the scene point location. An improved robustness to observation errors is achieved by defining the plurality of differences to penalize differences that includes the linear scaling as applied along the distance coordinate axes.

Abstract: A device determines positions of objects in a scene. The device obtains object detection data (ODD) which identifies the objects and locations of reference points of the objects in 2D images of the scene. The device processes the ODD to generate candidate association data (CAD) which associates pairs of objects between the images, computes estimated 3D positions in the scene for associated pairs of objects in the CAD, and performs clustering of the estimated positions. The device further generates, based on estimated 3D positions in one or more clusters, final association data (FAD) which associates one or more objects between the images, and computes one or more final 3D positions in the scene for one or more reference points of the one or more objects in the FAD. The final 3D position(s) represent the 3D position or the 3D pose of the respective object in the scene.

Abstract: A method for rescue support involves distributing transponder devices to individuals in a geographic region and causing the transponder devices to be activated. If a large-scale emergency situation occurs, a fleet of unmanned aerial vehicles are operated within the geographic region to receive distress signals transmitted by the activated transponder devices. A computer system processes the distress signals to obtain detection data for the individuals, the detection data comprising a location of the respective transponder device and a health status of the respective individual. The computer system generates, based on the location and health status of the respective individual, a prioritization chart for the geographic region, the prioritization chart being indicative of one or more sub-regions to be prioritized for a rescue operation.

Abstract: A tag (10) is configured to be attached to an object (12). The tag includes a sensor (24) configured to generate data indicative of movement of the tag in three dimensions. A control circuit (14) of the tag is configured to determine an amount of change in orientation of the tag over a predetermined period of time based on the data indicative of movement of the tag in three dimensions generated by the sensor; and to determine that the amount of change in orientation indicates that the tag is in one of an attached state relative to the object or is in a detached state relative to the object. A signal may be transmitted to a host computing device (28) if it is determined that the tag has become detached from the object.